You are here

Case Discussion: Lack of trial registration leads to new concerns about study conduct and ethical review/approval

Case Summary

Case 11-11

The editors of a journal noticed that a published article reported clinical trial results without mentioning trial registration. When asked to follow up, the authors submitted an unsatisfactory trial protocol and said the trial was unregistered. They then requested article withdrawal. The editors, however, said they planned to publish an Expression of Concern and asked for a copy of the approval letter from the ethics committee (institutional review board) mentioned in the article. At the same time, a member of the authors’ institution, who was not a co-author and wished to remain anonymous, informed the editors that the ethics committee named in the article did not exist at the institution.

After not hearing back from the authors, the editors added a Comment to the article alerting readers of non-adherence to clinical trial requirements. The authors later said that ethics approval had been granted by another institution. None of the authors seemed to be from that institution, and many issues including study setting, ethics approval, and participant recruitment remained unclear.

Question(s) for the COPE Forum

• Is retraction appropriate? The retraction notice would be updated if needed, after requesting an investigation at the authors’ institution.

Forum Advice and Follow-up

The Forum advised that the journal could issue an Expression of Concern. Most agreed on an institutional investigation followed by retraction if needed. However, some said that the lack of trial registration, unsatisfactory protocol, and possible lack of ethics approval already justified retraction.

The editors requested an institutional investigation, which found that the original trial had been conducted with appropriate ethical oversight. After being approved by a different ethics committee, the trial was done at the sponsoring company, with its employees as participants. In that country, the trial did not need registration because it was a post-marketing study of an approved drug. The published paper was actually an analysis of “samples” that the authors had received from the trial investigators.

The editors planned to publish an Expression of Concern to clarify the trial’s ethics approval, explain the company’s involvement, and state that although not required in that country, the trial should have been registered according to the criteria of the journal and the International Committee of Medical Journal Editors (ICMJE).

Case Discussion

This archived COPE Forum case is complex, and is categorised under several COPE Core Practices, which will be dealt with in turn below.

Human research ethics and complete reporting

A key COPE Core Practice in this case is Ethical Oversight, which states: “Ethical oversight should include, but is not limited to, policies on consent to publication, publication on vulnerable populations, ethical conduct of research using animals, ethical conduct of research using human subjects, handling confidential data and ethical business/marketing practices.”

The main problem with the published article in this case is that it is derived from a clinical trial whose details were not disclosed properly at submission or in the paper. The editors requested relevant documentary evidence such as the trial protocol and ethics approval letter, but it was not until the institutional investigation that an approved company-sponsored post-marketing trial (a type of Phase IV trial) was apparent. The study was based on “samples” from the trial that also had not been described properly. It is not known why the trial, trial sponsor, or the company donor of the materials were not clearly and completely disclosed. There are indeed still several unanswered questions.

It is unclear from the case details what the samples were, how they were selected, and whether their collection and analysis were included in the approved trial protocol and original informed consent agreement. In addition, we do not know if the published study represents the outcome-analysis part of the main trial, a secondary analysis (eg, studying secondary outcomes or an additional analysis of a subgroup of participants), or an independent post-hoc study using anonymised and de-identified samples. Additional ethical approval and participant consent, or a formal waiver of approval and consent, may still have been needed (retrospectively, if allowed). Some institutions require that human studies that are related to a main study approved elsewhere (including multicentre trials or certain contract research) receive at least local administrative approval.

However, this case presents an odd and unresolved situation where the named local institutional review board may not exist. The allegation was made by a whistleblower at the institution who wished to remain anonymous. The relevant COPE flowchart is “How to respond to whistleblowers when concerns are raised directly”. The journal should have requested the institution to follow up on that allegation and any associated misconduct.

Another relevant COPE Core Practice is Data and Reproducibility, which states: “Journals should include policies on data availability and encourage the use of reporting guidelines and registration of clinical trials and other study designs according to standard practice in their discipline.” This core practice is not referenced in the archived case, perhaps because the journal does seem to endorse the ICMJE’s requirement of trial registration in a public trials registry. It just did not implement or check that requirement.

The editors rightly pointed out that even though some countries may not legally require registration of post-marketing studies, they still need registration by journals using the ICMJE recommendations. The World Health Organization’s Q&A webpage on clinical trials states: “For the purposes of registration, a clinical trial is any research study that prospectively assigns human participants or groups of humans to one or more health-related interventions to evaluate the effects on health outcomes.… This definition includes Phase I to Phase IV trials.…Thus, early and late trials, trials of marketed or non-marketed products, randomized or non-randomized trials — all should be registered.” Although registration of observational study types (eg, cohort, case-control, and cross-sectional studies) is currently not compulsory, it is encouraged so as to maximise transparency and reproducibility.

Along with trial registration, complete and transparent reporting is important, and deficiencies in reporting should have been addressed in this case. Advice on publishing trials and their related studies is contained in the ICMJE recommendations and Good Publishing Practice (GPP3) guidelines. The appropriate EQUATOR Network reporting guideline could be used depending on the study design type. Specifically, secondary or related studies should cite the primary trial and disclose funding, role of the funder, and potential conflicts of interest. Redundant publication of data should be avoided and overlapping data or findings should be indicated as such. According to ICMJE, the clinical trial registration number should be stated in trial reports, and secondary or related studies should cite the primary study’s trial registration number. Non-registration and retrospective registration (if allowed) should be explained in journal articles.

Editorial office procedures

Administrative problems come to light in this case. Deficiencies in manuscript screening, editorial processing, peer review, and technical checking allowed an unclear study type with unclear methodology, sampling, and ethical review/approval to be published. Two core practices referenced in the reported case are as follows:

- Journal Management: “A well-described and implemented infrastructure is essential, including the business model, policies, processes and software for efficient running of an editorially independent journal, as well as the efficient management and training of editorial boards and editorial and publishing staff.”

- Peer Review Processes: “All peer review processes must be transparently described and well managed. Journals should provide training for editors and reviewers and have policies on diverse aspects of peer review, especially with respect to adoption of appropriate models of review and processes for handling conflicts of interest, appeals and disputes that may arise in peer review.”

Although the editors said the journal requires registration and a study protocol for trials, and later asked about these along with proof of ethical approval, these items were apparently not requested at the initial manuscript submission/screening step. The journal would do well to comprehensively review its internal procedures and implement measures to prevent future problems while improving the quality and integrity of its publication.

Journals should ensure they have adequate submission systems and processes to collect information such as study design type, declaration of secondary or overlapping studies, complete reporting (eg, via EQUATOR Network reporting guideline checklists), trial registration, and required paperwork, approvals, and declarations. They should also ensure that author instructions in general, and instructions in the submission platform in particular, are clear, up-to-date, and complete. Steps need to be taken, however, to avoid allowing authors to automatically tick/check compulsory checklist items: appropriate evidence should be uploaded too.

Office staff need to be trained to conduct necessary submission-checking procedures and confirm supplied evidence. Peer reviewers and editors could be given checklists and instructions to verify study design types and comment on missing sections according to reporting guidelines. All journal personnel could be specifically asked to spot ethical problems during submission and review. Problems could then be handled according to the COPE flowchart on “What to do if you suspect an ethical problem with a submitted manuscript”. The office should have relevant guidelines in place, as well as a designated person or team who responds to complaints, ethical issues, and allegations of misconduct.

Journals can periodically check they are using and referring to the latest guidelines on best publishing practice (eg, from ICMJE and COPE) and revise their own guidelines for authors and reviewers. For example, authors should be asked to include specific information to do with ethics and transparency in a manuscript’s endmatter. Standard declarations include authorship contributions, non-author contributions and donors of materials in the acknowledgements, disclosure of funding (including contract research), role of the funder in any of the study or reporting, ethical approval and consent, whether all authors had access to the study data, and data accessibility statements. Office staff, technical editors, or production staff could be trained to check such endmatter.

Relevant COPE guidance for editorial office management is contained in the document titled “General Approach to Publication Ethics for the Editorial Office”. Offices of non-COPE journals can align their policies and procedures with the 10 COPE Core Practices and link to COPE resources (but not reproduce or copy them or misuse the COPE logo). They can use the core practices as a basis for adhering to the 16 Principles of Transparency and Best Practice in Scholarly Publishing in preparation for future application for COPE membership. They can also freely access the Introduction to Publication Ethics module of the COPE E-learning course.

Offices of COPE-member journals can use their customised member logo and should refer to the 10 COPE Core Practices and resources archived under each core practice to create or revise their own codes of conduct. They should also use the latest COPE Journal Audit (member-only access) to check and update their manuscript handling and ethics policies and procedures. The audit document also serves as a handy syllabus guide for training office staff in publication ethics. In addition, office staff should be encouraged to complete all modules of the COPE E-Learning course (member-only access). To keep in touch with COPE, journals/publishers should inform COPE of any updates for their chosen COPE representative (see, or email [email protected]).

Correcting the scholarly record

The fourth COPE Core Practice listed in the archived case is Post-publication Discussions and Corrections, which states: “Journals must allow debate post publication either on their site, through letters to the editor, or on an external moderated site, such as PubPeer. They must have mechanisms for correcting, revising or retracting articles after publication.”

Transparency and accuracy in reporting and full disclosure of ethical aspects should be addressed in any post-publication remedial action. The editors in this case initially planned to publish an Expression of Concern but decided to opt for a Comment as an initial alert about non-compliance to the journal’s clinical trial requirements. Comments are unofficially but widely regarded as a softer alternative or potential prequel to an Expression of Concern. The editors’ initial chosen action of a post-publication comment contrasts with the withdrawal requested by the authors. However, withdrawal (removal) of a published article is reserved only for legal or health-risk reasons.

After the institutional investigation, an Expression of Concern was finally planned, to explain the original sponsored trial and state that it should have been publicly registered. Expressions of Concern are used to alert readers of a possible concern about the integrity of an article. They may be issued before, during, after, or in lieu of an investigation, and are still a relatively new type of post-publication notice that is being discussed by COPE.

By planning to issue an Expression of Concern, the editors considered the case closed. However, if the editors were satisfied with all the conclusions from the institutional investigation, they should have also made necessary factual clarifications as formal corrections in the article’s text, at least about the study’s relationship to the original trial, ethical approval and consent, funding, conflicts of interest, and acknowledgements. We are bound to assume that the ethical approval and participant consent for the published study were adequately accounted for by the original trial and explained as such in the Expression of Concern. Otherwise, if the ethical approval and consent details of the study were still unclear, the Expression of Concern should have stated this and the matter should have been investigated further, leaving open the possibility for future retraction.

Trevor Lane on behalf of the COPE Education Subcommittee